C / C++ / MFC

12-Jun-10 5:41

I need some code for a template

matrix gauss_jordan(void) { // Gauss-Jordan matrix inversion
assert(this->data[].size() == this->data.size()); // Must be a square
};

I have not seen much of help so far

I do have a code to put a matrix into row echelon form

http://www.contract-developer.tk

modified on Saturday, June 12, 2010 3:59 PM

Re: gauss jordan

Trollslayer13-Jun-10 0:14

Trollslayer

13-Jun-10 0:14

That is because the message boards are about answering questions and providing guidance for those who are doing the work.
You are asking for someone to do the work for you.
When you have made some progress then by all means ask about specific issues or points.

Join the cool kids - Come fold with us[^]

Re: gauss jordan

ThatsAlok13-Jun-10 18:24

ThatsAlok

13-Jun-10 18:24

VeganFanatic wrote:
assert(this->data[].size() == this->data.size()); // Must be a square

what is data type of data!

cheers,
Alok Gupta
VC Forum Q&A :- I/IV
Support CRY- Child Relief and You

Re: gauss jordan

VeganFanatic13-Jun-10 18:28

About OnFilePrint and OnPreparePrinting question.

13-Jun-10 18:28

I make everything a template. This way its agnostic to the underlying data to a point.

I have templates for vector and matrix already and I can now use vector equations, so when I made my matrix I leveraged that capability a lot. Make the work a lot easier.

I finally have LU decomposition, I wanted other algorithms in the library to be used in special cases.

http://www.contract-developer.tk

wangningyu11-Jun-10 21:53

wangningyu

11-Jun-10 21:53

Hello everybody!

I create the class used to show some graphics, like this:

CMyElemView >> CZoomView >> CScrollView >> CView

when I called this,but is empty .

void CMyElemView ::PrintPreview()
{
	CScrollView::OnFilePrintPreview();
}
void CMyElemView ::Print()
{
	CScrollView::OnFilePrint();
}

otherwise , I created other class,it used to show some text:

CMyTextView >> CTextBaseView >> CBCGPEditView >> CView

This can be successed to show the text, but the CMyElementView doesn't.

What cause is this error generally ?

Templated array in class

grusky11-Jun-10 14:27

grusky

11-Jun-10 14:27

I have created a class and a tamplate to create an array with a size based on two of the parameters. What I am looking to do is populate the years[0] with the start parameters value and continue to year[stop]= the stop parameters value.

#include<iostream><br />
using namespace std;<br />
<br />
template <class T, int start, int stop><br />
class MyArray {<br />
	T test [stop-start];	<br />
};<br />
<br />
int main () {<br />
MyArray <int, 1790, 1820> years; <br />
return 0;<br />
}<br />

so... MyArray <int, 1790, 1820> years; would create an array with 30 values starting at array[0] = 1790 and ending with array[30]=1820. Is this possible?

Rick York11-Jun-10 15:46

Rick York

11-Jun-10 15:46

To begin with, I think you want stop-start+1 because 1790 to 1820 is actually 31 years. Correspondingly, if you passed in 1 and 10 you should have 10 items and not 9, right ?

Anyway, no, I don't think you can do exactly what you wrote there. Have you seen what the compiler thinks about that ? It will probably tell you that the array size must be a constant expression. I think you need to use new to make the array of variable size.

Nemanja Trifunovic11-Jun-10 16:35

Nemanja Trifunovic

11-Jun-10 16:35

As long as the start and stop values are compile-time constants, I am pretty sure you can use Template Metaprogramming[^] to populate the array.

utf8-cpp

«_Superman_»11-Jun-10 20:51

«_Superman_»

11-Jun-10 20:51

I recommend you use the std::vector class for this.
You can do the following -

std::vector<T> myArray;
myArray.reserve(stop - start + 1);

«_Superman_»

I love work. It gives me something to do between weekends.

Microsoft MVP (Visual C++)

Polymorphism in C

Niklas L11-Jun-10 22:40

Niklas L

11-Jun-10 22:40

I think you meant vector::resize().

---

IMO using the std::vector does not add any value in this case. In fact it's the other way around.

* It will allocate memory from the heap, whether you have an automatic instance or not. Possibly affecting performance and fragmentation. (Not a problem for most application, but still)
* There is no added bounds checking. It is just as good (or bad) as the native array.
* Memory management is provided for you in both cases.

Use of std::vector might be useful if:
* You work with years spanning over millions of years. Otherwise it might consume your stack in no-time.
* Maybe the compiler/linker can optimize your class due to the fact that all instances (with the same type parameter) have the same size. (sizeof(std::vector<int>) is not affected by the number of elements while sizeof(int[N]) is.) Unfortunately, my knowledge of linker optimization techniques is not what it should be so I really don't know. But I wouldn't mind knowing. Smile | :)

home

Niklas L11-Jun-10 21:56

Niklas L

11-Jun-10 21:56

template <class T, int start, int stop>
class MyArray {
	T test [stop-start+1];
public:
	MyArray()
	{
		for (int i = 0; i < (stop - start + 1); ++i)
			test[i] = start + i;
	}
};

MyArray<int, 1999, 2010> arr;

home

Aescleal12-Jun-10 8:12

Aescleal

12-Jun-10 8:12

Here's a quick counter question... Do you actually need an array?

The reason this springs to mind is that you've made the array private so no one outside of your interface can use it. Well, apart from using some disgusting tricks that you should burn in a personal hell for.

So if your final class interface is going to be something like:

template <class T, int first, int last>
class year_range
{
    public:
        T nth_year_in_range( int n );
};

you could actually eliminate the array wholesale:

T nth_year_in_range( int n )
{
    return T( start + n );
}

Of course if you need the array for something else then this won't work.

Cheers,

Ash

matrix operator =

VeganFanatic11-Jun-10 12:13

11-Jun-10 12:13

matrix operator=(const matrix &that) {
matrix ret;
ret.resize(this->data.size(), this->data[0].size());
for (int i = 0; i < that.data.size(); i++)
for (int j; j > that.data[0].size(); j++)
ret[i][j] = that[i][j];
return ret;
};

Seems to be no good, what am I doing wrong

http://www.contract-developer.tk

image processing toolkits | batch image processing

C^hris L^osinger11-Jun-10 15:05

C^hris L^osinger

11-Jun-10 15:05

are you getting an error message of any kind ?

VeganFanatic11-Jun-10 15:39

11-Jun-10 15:39

Claims there are no operators available, to support the abstract vector algebra.

Error 2 error C2679: binary '-' : no operator found which takes a right-hand operand of type 'aa::vector<base>' (or there is no acceptable conversion) matrix.h 114 1 fibonacci

http://www.contract-developer.tk

VeganFanatic11-Jun-10 15:42

11-Jun-10 15:42

vector operator-(const vector *that) { // parallelogram law
vector ret;
assert(this->data.size() == that.data.size());
ret.resize(this->data.size());
#pragma omp parallel for
for (int i = 0; i < this->data.size(); i++)
ret.data[i] = this->data[i] - that.data[i];
return ret;
};

http://www.contract-developer.tk